3 ms·

"If we choose music like we sometimes choose technology, we would all be listening to Justin Bieber. The cow paths aren’t justification enough to use a technolo

"If we choose music like we sometimes choose technology, we would all be listening to Justin Bieber. The cow paths aren’t justification enough to use a technology or adopt a pattern." I wholeheartedly agree with the quote. Almost each technology wave takes its toll. Remember how everybody got excited about NoSQL? It's not that this technology hardly matters nowadays - it's just clear it's relevant only in certain contexts, just like everything else. Yesterday it was about "microservices", today it's "serverless". It's almost as if people were bending over backwards so that the technology currently en vogue matches their problem/solution scenario. As is people were afraid of thinking independently and challenging what happens to be the hype of the moment.

BinaryIdiot 9y ago
> Remember how everybody got excited about NoSQL? It's one of the downsides of being a contractor; your new client thinks NoSQL is the next best thing and if you want the job you have to use NoSQL, even if it makes zero sense in your use case (and yes you always argue with the client it just rarely works if they have the money to throw at the issue). When the NoSQL craze hit I stopped using RDMS entirely for about 6 years. It was never my choice and only in one case do I think it was the correct choice.
- dvfjsdhgfv 9y ago
  For me the worst was the XML craze. Suddenly everyone was talking about XML as if it was a game changer. One major database vendor announced an XML database. New standards like XSLT and Xpath proliferated. XHTML seemed to be on the way to replace HTML... When you look at it now, each of these technologies has some use, and the "XML revolution" contributed to some useful developments, but the exaggerations on the way were egregious. What is interesting is that many people were aware of the shortcomings of the new technology from the very beginning. For example, many noted that XML is too verbose for both machine-to-machine communication as well as editing by humans. It's interesting to note the same phenomena in each technology cycle.
alexasmyths 9y ago
Everything goes through the hype-cycle ie 'trough of disillusionment'. 'Serverless' is one that is here to stay for sure. Nobody likes having to manage 'instances'. There's a gold mine for anyone who can reliably make something that 'just runs your code at scale' - in a normal format, without a lot of weird apis or behaviours. AWS Lambda's are 'almost it' but there's a few caveats that make them annoying enough that they are not the solution. NoSQL are limited mostly by: A) lack of 'joins' in queries and B) too many query languages. I suggest both of those problems are solvable, in which case they might be even better than RDBMS for many apps. Also - many people just don't know how to 'data model' for them. Or maybe RDBMS will come a long with an easy 'flex table' solution and solve the problem noSQL was trying to solve and make it relevant. (Who knows, maybe they have? I don't pay close attention to that community). As for Justin Bieber - well - there's 'no advantage' to everyone listening to 'Justin Bieber' - but because technology has to 'work together' - there is an existential advantage when we 'all use the same stuff'. 1 solid tool that everyone uses is probably better than 100 different tools, each of them a little bit better in different ways, and nobody on the same page. Finally - the bleeding edge is important though, it's the tip of innovation where new stuff happens :)
marcosdumay 9y ago
Honestly, we have to bend everything out of shape to make any functional software for any domain, using any technology. It is just natural that people have problems identifying wether stuff suits their problem or not.

BinaryIdiot 9y ago

> Remember how everybody got excited about NoSQL? It's one of the downsides of being a contractor; your new client thinks NoSQL is the next best thing and if you want the job you have to use NoSQL, even if it makes zero sense in your use case (and yes you always argue with the client it just rarely works if they have the money to throw at the issue). When the NoSQL craze hit I stopped using RDMS entirely for about 6 years. It was never my choice and only in one case do I think it was the correct choice.

dvfjsdhgfv 9y ago
For me the worst was the XML craze. Suddenly everyone was talking about XML as if it was a game changer. One major database vendor announced an XML database. New standards like XSLT and Xpath proliferated. XHTML seemed to be on the way to replace HTML... When you look at it now, each of these technologies has some use, and the "XML revolution" contributed to some useful developments, but the exaggerations on the way were egregious. What is interesting is that many people were aware of the shortcomings of the new technology from the very beginning. For example, many noted that XML is too verbose for both machine-to-machine communication as well as editing by humans. It's interesting to note the same phenomena in each technology cycle.

alexasmyths 9y ago

Everything goes through the hype-cycle ie 'trough of disillusionment'. 'Serverless' is one that is here to stay for sure. Nobody likes having to manage 'instances'. There's a gold mine for anyone who can reliably make something that 'just runs your code at scale' - in a normal format, without a lot of weird apis or behaviours. AWS Lambda's are 'almost it' but there's a few caveats that make them annoying enough that they are not the solution. NoSQL are limited mostly by: A) lack of 'joins' in queries and B) too many query languages. I suggest both of those problems are solvable, in which case they might be even better than RDBMS for many apps. Also - many people just don't know how to 'data model' for them. Or maybe RDBMS will come a long with an easy 'flex table' solution and solve the problem noSQL was trying to solve and make it relevant. (Who knows, maybe they have? I don't pay close attention to that community). As for Justin Bieber - well - there's 'no advantage' to everyone listening to 'Justin Bieber' - but because technology has to 'work together' - there is an existential advantage when we 'all use the same stuff'. 1 solid tool that everyone uses is probably better than 100 different tools, each of them a little bit better in different ways, and nobody on the same page. Finally - the bleeding edge is important though, it's the tip of innovation where new stuff happens :)

marcosdumay 9y ago

Honestly, we have to bend everything out of shape to make any functional software for any domain, using any technology. It is just natural that people have problems identifying wether stuff suits their problem or not.